RAMAN Sukumar's book on elephants begins with a bull with a headache. We meet Biligiri in the first paragraph: he is an adolescent wild male elephant, "confused, like other sixteen year olds" (sic), with "a dull ache in his temples". He thrashes about in water, and does some pretty strange things as elephants go. The reason for Biligiri's behaviour, Sukumar would have us believe, is because for the first time he is in "musth".

Old wives' tales had it that the fluid which runs out of the temporal gland when a elephant is in musth causes a painful congestion in the head, leading the pachyderms to behave aggressively. But elephant science has come a long way since then, and the musth condition is now attributed to high levels of testosterone in the body leading, boosting aggression. No one has an elephant's word on whether an overdose of good old testosterone in the body causes a headache.

While Sukumar addresses an important aspect of elephant behaviour in his study -- why they are driven to raid crops -- the information he provides in his book is often ambiguous. He tells us, for example, that ultimately, elephants raid crops as part of a high-risk, high-gain strategy, because evolution has shaped feeding behaviour so that the animal seeks the best available food resource. This would imply anything edible that comes their way -- whether growing naturally or human-grown -- whereby generations of elephants have found that crops are more nutritious, that eating them ensures better reproductive success, and that the benefits they accrue from eating crops are worth the risk they have to take getting to them.

Bulls, says Sukumar, are more willing to take this risk because a bigger and healthier male would "dominate the mating game". But turn a page and he tells us that agriculture has not been around long enough for crop raiding behaviour to have been shaped through selection, and that he is "not suggesting that crop raiding bull elephants enjoy better reproductive success than those which do not raid crops".

Sukumar does seem to be pushing the natural selection theory to absurd lengths. First, we have elephants raiding crops because of natural selection, and then he goes on to imagine that natural selection favoured fire-setting behaviour in primitive man. Hunter-gatherers or shifting-cultivators in small settlements in the wilderness 20,000 years ago would be threatened by accumulated biomass around their settlements, says Sukumar, because it could catch fire and destroy everything in its path. But if clans set periodic "harmless ground fires" to clear their surroundings, they would reduce the likelihood of an enemy clan igniting their settlement, or a natural fire causing excessive destruction. So, he concludes, fire setting behaviour would have been favoured by natural selection.

Sukumar falls easily into the trap of importing research without looking at its implications and applicability in India. Take, for example, his extension of Graeme Caughley's model of the relationship between elephants and trees in East Africa: Caughley, a population ecologist, had suggested that elephants and trees are locked in a cyclic relationship. An overabundance of elephants would lead to a crash in the vegetation and a shortage of food, and thus a crash in elephant populations. This crash, in turn, would allow the vegetation to recover, and the elephant population would increase again. In the Luanghwa Valley of Zambia, Caughley suggested, the period between successive peaks or troughs in the population density of trees and elephants was about 200 years.

Sukumar extends Caughley's basic model interaction to all kinds of elephant habitats without taking into account that in Africa, troughs and peaks are possible because adjoining habitat exists and allow time for the recolonising and recovery of damaged habitat. But in India, where elephants are trapped in small islands, there can no longer be "peaks" in habitat quality after a crash.

While extending Caughley's model to rainforests, Sukumar says that since elephants cannot push over huge trees, and many of the plants in the understorey are unpalatable because they contain toxic substances, rainforests support a low number of elephants. So, he says authoritatively, elephants and rainforests exist in near stable equilibrium.

There is a problem with this theory. Although the forest cover may remain unchanged because of the large trees and unpalatable undergrowth, what's to say that the elephant food plants and elephant numbers in rainforests are not going through cyclic changes? The actual forest cover is no indication of the interaction between plant and elephant. Besides, no study has been carried out to prove that elephants and rainforest exist in equilibrium.

Finally, the book's title is unjustified. Nowhere does the reader get the impression that the author has an intimate knowledge of elephants. Iain and Oria Douglas-Hamilton based their book, Among the Elephants, on a 5-year study in Manyara, Tanzania. Cynthia Moss' earlier book, Elephant Memories, was written after 13 years of field work in Amboseli, Kenya. Both books talk of individual clans and the behaviour and the daily life of individual elephants.

Sukumar's distance from his subject is perhaps in part because he spent the greater part of his decade in Bangalore, effectively spending only 3 years with the Asian elephant.